How andtsa/bsc_thesis works
Project Type and Purpose
• It is a research-oriented, multi-crate Rust application that:
• Implements core ranking definitions and Kendall’s τ computations (lib/)
• Provides two solvers (graph-based and brute-force) as CLI tools (solver/)
• Verifies solver outputs against reference data (verifier/)
• Runs end-to-end evaluations over CSVs (eval/)
• Generates R-based plots of results (plots/)
• Holds reference solution sets (solutions/).
Main Components
• Core Library (lib/)
• Solver CLI (solver/)
• Graph-based and brute-force algorithms
• Reference Solutions (solutions/)
• Verifier CLI (verifier/)
• Evaluation Tool (eval/)
• Plotting Scripts (plots/)
Interactions & Data Flows
a. Input data (rankings CSV) → Solver CLI
b. Solver CLI uses Core Library to compute τmin/τmax → writes solver output (CSV)
c. Reference Solutions crate holds “ground truth” solution data → consumed by Verifier
d. Verifier CLI reads solver output + reference solutions → emits verification report
e. Eval tool (eval/) orchestrates repeated solver+verifier runs over datasets → aggregates results
f. Plotting scripts (plots/*.R) read aggregated results (CSV) → output figures for the paper
Architectural Patterns & Design Principles
• Modular monorepo: each Cargo crate is a reusable module.
• CLI-Library separation: business logic in lib/, interfaces in solver/ and verifier/.
• Data-flow pipeline: clear stages (solve → verify → evaluate → plot).
• Single-responsibility per crate.
Key Technologies & Frameworks
• Rust/Cargo for all core components
• CSV I/O (serde, csv crates)
• Graph library for graph-based algorithm (in solver/)
• R for statistical plotting (in plots/)